Large site with content silo's - best practice for deep indexing silo content

[DrewProZ]

Thanks in advance for any advice/links/discussion.  This honestly might be a scenario where we need to do some A/B testing.

We have a massive (5 Million) content silo that is the basis for our long tail search strategy.  Organic search traffic hits our individual "product" pages and we've divided our silo with a parent category & then secondarily with a field (so we can cross link to other content silo's using the same parent/field categorizations).

We don't anticipate, nor expect to have top level category pages receive organic traffic - most people are searching for the individual/specific product (long tail).  We're not trying to rank or get traffic for searches of all products in "category X" and others are competing and spending a lot in that area (head).

The intent/purpose of the site structure/taxonomy is to more easily enable bots/crawlers to get deeper into our content silos.  We've built the page for humans, but included link structure/taxonomy to assist crawlers.

So here's my question on best practices.  How to handle categories with 1,000+ pages/pagination.  With our most popular product categories, there might be 100,000's products in one category.  My top level hub page for a category looks like www.mysite/categoryA and the page build is showing 50 products and then pagination from 1-1000+.

Currently we're using rel=next for pagination and for pages like www.mysite/categoryA?page=6 we make it reference itself as canonical (not the first/top page www.mysite/categoryA).

Our goal is deep crawl/indexation of our silo.  I use ScreamingFrog and SEOMoz campaign crawl to sample (site takes a week+ to fully crawl) and with each of these tools it "looks" like crawlers have gotten a bit "bogged down" with large categories with tons of pagination.  For example rather than crawl multiple categories or fields to get to multiple product pages, some bots will hit all 1,000 (rel=next) pages of a single category. I don't want to waste crawl budget going through 1,000 pages of a single category, versus discovering/crawling more categories.

I can't seem to find a consensus as to how to approach the issue.  I can't have a page that lists "all" - there's just too much, so we're going to need pagination.  I'm not worried about category pagination pages cannibalizing traffic as I don't expect any (should I make pages 2-1,000) noindex and canonically reference the main/first page in the category?).  Should I worry about crawlers going deep in pagination among 1 category versus getting to more top level categories?

Thanks!